Live data from Hacker News

Internet voting is insecure and should not be used in public elections

blog.citp.princeton.edu

41–50 of 532 posts

Re: Internet voting is insecure and should not be used in public elections

#41
post #5

The most important feature of public elections is trust. Efficiency is one of the least important feature. When we moved away from paper voting with public oversight of counting to electronic voting we significantly deteriorated trust, we made it significantly easier for a hostile government to fake votes, all for marginal improvements in efficiency which don't actually matter. Moving to internet voting will further…

The US overwhelmingly uses paper voting (often paired with electronic tabulation). We can't "move back", it's where we are.

Electronic tabulation introduces little risk when the ballots are paper.

Re: Internet voting is insecure and should not be used in public elections

#42
post #13

Earlier quoted context omitted.

Isn't a vote being verifiably tied to a person actually a bad thing? Then you can actually check what e.g. your wife or kids voted for and punish them if they vote wrong. Or get people to pay for votes, but doing that at scale is obviously hard. Maybe this isn't what you meant by verifiable, but there are systems with this property and they are bad.

Verifiable in this context means I can verify my vote was tallied correctly.

That would also mean someone could force you to show who/what you voted for.

Re: Internet voting is insecure and should not be used in public elections

#43

Earlier quoted context omitted.

Internet voting needs to be anonymous and non demonstrable. Internet money needs to be the opposite, and reversible through the courts.

It can't be anonymous. There has to be some form of IDV to ensure it is a registered voter.

The vote needs to be anonymous, not the registration + checkin process.

Re: Internet voting is insecure and should not be used in public elections

#44
post #13

Earlier quoted context omitted.

Isn't a vote being verifiably tied to a person actually a bad thing? Then you can actually check what e.g. your wife or kids voted for and punish them if they vote wrong. Or get people to pay for votes, but doing that at scale is obviously hard. Maybe this isn't what you meant by verifiable, but there are systems with this property and they are bad.

Verifiable in this context means I can verify my vote was tallied correctly.

Receipt-freeness (i.e., a secret ballot) is usually the desired property. Yes, a lot of people like you state they desire verifiable votes. But that's where you need to respond to the points the person above you is making: how is such a system not also susceptible to coercion and bribery?

(However you would verify your vote, imagine the person who is coercing you is just standing over your shoulder with threat of force. An example might be an abusive husband who does not want to allow their wife to vote freely/against him. A briber might simply force you to allow them to look over your shoulder before they'll pay you off.)

Vs. paper ballots in a polling place: a coercer would not be permitted in the poll booth with me. I get to vote, and when I leave, … I can tell them whatever, but it does not need to match my vote. It utterly defeats bribery, as the briber has no way to verify that I'm doing what they way.

Re: Internet voting is insecure and should not be used in public elections

#45
The thing about paper ballots is that the ways to cheat with them are well-known ("finding" ballots in the trunk of a car, "losing" ballot boxes on the way to the counting center, counting the ballots behind locked doors with observers not present, and so on), and have been well known for centuries. So the counters to them (ballot boxes sealed with an official seal once full, only sealed ballot boxes will be opened and counted, neutral observers present at all times when ballot boxes are being transported and/or counted, and so on) are also well-known. If those anti-cheating counters are in place, that gives you quite a lot of trust in the results. And if observers get thrown out and then ballot counting continues behind closed doors, you can have a reasonable suspicion that cheating is going on, and can make a stink and demand a redo of the vote.

With Internet voting, the ways to cheat are not all that well-known among the general population, and even among an audience like HN I bet we couldn't come up with all the ways to cheat. (That's not a challenge!) So there's going to be fundamentally less trust in the election process than with paper ballots, even if the Internet-voting system was actually made completely secure. (And I'm not persuaded it can be made completely secure, given that secret ballots are a fundamental requirement of the process).

So yes, paper ballots are very much the way to go.

Re: Internet voting is insecure and should not be used in public elections

#46

Earlier quoted context omitted.

Internet voting needs to be anonymous and non demonstrable. Internet money needs to be the opposite, and reversible through the courts.

It can't be anonymous. There has to be some form of IDV to ensure it is a registered voter.

The ballot has to be anonymous, or unable to be tied back to the voter once cast. It’s a hard requirement for a variety of reasons

Re: Internet voting is insecure and should not be used in public elections

#48

Earlier quoted context omitted.

Internet voting needs to be anonymous and non demonstrable. Internet money needs to be the opposite, and reversible through the courts.

It can't be anonymous. There has to be some form of IDV to ensure it is a registered voter.

Indeed, many people now get a erroneous covid tax-relief refund bill for not qualifying for a program they never signed up for in the first place.

One local scammer made off with a $5m government refund for a fraudulent business tax filing. You can't make this stuff up if you tried...

At some point, one is just amazed at the size of the cons people pull online. =3

Re: Internet voting is insecure and should not be used in public elections

#49
post #45

The thing about paper ballots is that the ways to cheat with them are well-known ("finding" ballots in the trunk of a car, "losing" ballot boxes on the way to the counting center, counting the ballots behind locked doors with observers not present, and so on), and have been well known for centuries. So the counters to them (ballot boxes sealed with an official seal once full, only sealed ballot boxes will be opened a…

Oh, and if the election is on something so polarizing that there are no "neutral" observers, then rather than neutral observers you can have observers from both (or all) parties/sides present, with cameras rolling, while the counting is going on.

Re: Internet voting is insecure and should not be used in public elections

#50
post #13

I believe the piece we're missing is the government (citizen?) service which issues (manages, replaces, revokes) constituents' cryptographic tokens for use with such things. Then our voting systems could be electronic, secure, open, verifiable, and mostly private; assuming effective oversight / this organization does not issue fraudulent tokens or leak keys or identities (big assumption, but I don't think it's imposs…

Isn't a vote being verifiably tied to a person actually a bad thing? Then you can actually check what e.g. your wife or kids voted for and punish them if they vote wrong. Or get people to pay for votes, but doing that at scale is obviously hard. Maybe this isn't what you meant by verifiable, but there are systems with this property and they are bad.

The property you are talking about is generally called "deniability" in the literature, whereas the GP is talking "verifiability" ie. being able to verify your own vote is cast correctly. They are both valuable, sometimes mutually exclusive, but not necessarily, see eg. https://petsymposium.org/popets/2024/popets-2024-0021.pdf
Post reply on HN